Margolin et al., 2004b
A Margolin, N Banerjee, I Nemenman, and A Califano. Reverse engineering of yeast transcriptional network using the ARACNE algorithm. Unpublished manuscript, 2004. PDF.
- Cellular phenotypes are determined by dynamical activity of networks of co-regulated genes. Elucidating such networks is crucial for the understanding of normal cell physiology as well as for the dissection of complex pathologic phenotypes. Recently we have shown that ARACNE, a novel information-theoretic algorithm for reverse engineering of transcriptional networks using microarray data, holds significant promise for the genome-wide analysis of mammalian networks, which had never been performed in silico. In this paper we present the application of ARACNE to reverse engineering of transcriptional networks in the yeast Saccharomyces cerevisiae. This provides another platform for further comparisons of the new method to a variety of established ones, which have been extensively used to analyze this important model organism. Moreover, it provides an additional genome-wide interactome for the yeast using a method that is shown to produce very few false positive interactions, both in vivo and in silico. Additionally, we investigate the global topological properties of the reconstructed networks and determine that the scale-free structure suggested by existing models should be taken cautiously. Finally, analysis of the ARN1 sub-network, which has also been investigated using Bayesian Networks, shows that ARACNE is able to distinguish the key regulatory elements of this pathway from within a large cluster of co-regulated genes.